The spread of viruses and worms has severe implications on the performance of virtually any network. Known methods for stopping the propagation of malicious code generally rely on anti-virus signature recognition to prevent hosts from being infected. Unfortunately, the latency between the introduction of a new virus into a network and the implementation/distribution of a patch can be significant. Within this period, a network can be crippled by the abnormally high rate of data traffic generated by infected hosts.